Propaganda

What’s, why’s, and what-do-we-do-about-it’s

Georgiy Syunyaev

g.syunyaev@vanderbilt.edu

October 29, 2025

Introduction

  • Personal history

    • Geography: 🇷🇺 \(\Rightarrow\) 🇺🇸 \(\Rightarrow\) 🇩🇪 \(\Rightarrow\) 🇺🇸
    • Disciplines: Math in School \(\Rightarrow\) Econ (Under)Grad \(\Rightarrow\) PhD in Political Science \(\Rightarrow\) Post-Doc (Methods) \(\Rightarrow\) Assistant Prof at Vandy

  • I study when and why media persuades in autocracies (esp. Russia) and beyond

    • How citizens process media and elite messaging, or use internet for politics
    • What audiences “get” from state media
    • Can shifting attention change media perceptions and downstream beliefs?

  • I teach

    • Politics of Ukraine & Russia; Research Design/Methods
    • Considering a course on propaganda & misinformation

What is propaganda?



  • Strategic control of information by the state to shape citizens beliefs and behavior

  • Includes spin (framing/priming) and censorship/self-censorship (Roberts 2018, Princeton)

    • Flooding/jamming hides sensitive info (Paul & Matthews 2016, RAND)
    • Misinformation and conspiracies can be tools of propaganda

  • From hard propaganda (ideology, fear) to soft propaganda (competence, entertainment) (Guriev & Treisman 2022, Princeton)

    • Soft propaganda masks under non-political content (e.g. TV shows or movies)

Why do governments use it?



  • Persuasion: win support for policies and regime (Pan, Shao & Xu 2022, PSRM; Syunyaev 2025, WP)
  • Distraction & demobilization: raise apathy; lower perceived agency (Huang 2015, CP; Tertytchnaya & Lankina 2020, JOP)
  • Identity & nation-building: cultivate nationalism and hegemonic worldviews (Mattingly & Yao 2022, CPS)
  • Signaling power & unity: credible strength and competence cues (Guriev & Treisman 2022, Princeton)

Persuasion vs. domination



  • Persuasive content must be plausible and low-key

    • Mix truth with fiction; avoid obvious lies. “Best when least obvious.”

  • Dominating tactics crowd out alternatives and potentially lead to apathy (especially, flooding or bots)

  • These tools don’t always complement each other and are hard to reverse

Is propaganda effective?



  • Often Yes! both at micro and aggregate levels (Peisakhin & Rozenas 2018, AJPS; Mattingly & Yao 2022, CPS; Krishnarajan & Tolstrup 2023, Science Advances)
  • Can be moderated: effects often vary across subgroups (education, partisanship, critical thinking); context matters as well (Selb & Munzert 2018, APSR; Bleck & Michelitch 2017, JOP; Adena et al. 2015, AER) \(\Rightarrow\) what are the population effects?

  • Other paths to “success”:

    • Induce cynicism and disengagement \(\Rightarrow\) reduce agency (Alyukov 2022, NHB)
    • Build mass tolerance for/fear of repression/censorship (Yang 2025, JOP; Tertytchnaya & Lankina 2020, JOP)

Evidence

  • 1930s: Radio Propagation in Weimar Germany and its Effects on Voting for Nazi Party (Adena et al. 2015, AER)

  • 1936-42: Father Coughlin’s Effects on FDR Votes and Sales of US War Bonds (Wang 2021, JPE)

  • 1943-45: Italians’ Exposure to BBC Radio and Resistance to Nazi Occupation during WWII (Gagliarducci et al. 2021, )

  • 1994: Exposure to Hutu Nationalist Radio and Participation in Rwandan Genocide (Yanagizawa-Drott 2014, QJE)

  • 2012: Random Radio Distribution and Exposure to Radio during the Mali Coup (Bleck & Michelitch 2017, JOP)

  • 2014: Exposure to Russian TV and Ukrainian Elections in 2014 (Peisakhin & Rozenas 2018, AJPS)

  • 2018: Expansion of Transmission of Independent Radio Station in Tanzania Green et al. 2024, JOP

Precincts within radius of Nazi controlled radio broadcasts were more likely to vote for Hitler, join the Nazi Party and engage in anti-Jewish deportations prior to 1942. But the effects were moderated by past anti-Semitic attitudes.

Leveraging variation in topography (but controlling for distance) identifies negative (-2.4 p.p.) effects on FDR votes which survived even after broadcast going away from public

Using sunspot activity to approximate variation in BBC radio reception finds important role of radio in motivating resistance but no long-lasting anti-Nazi effects

Topographic variation in exposure to RTLM responsible for roughly 10% of killings, especially from violence requiring coordination

Radio exposure boosted national identity but did not elevate explicit support for the junta

Areas with higher cross-border exposure to Russian state media had higher support for pro-Russian parties, but no effect on turnout

Effects on political interest and knowledge about domestic politics but sporadic changes in attitudes on a range of gender issues (covered by radio)

Why does it work?




  • Inattention: System 1 (reasoning based on cues, e.g. partisanship) vs System 2 (relying on reflective and attentive reasoning)
  • Motivated reasoning: interpreting new information to reach certain goal (also can explain consumption!)

When does it (not) work?

  • Content: attribution and framing are effective (Pan, Shao & Xu 2022, PSRM; Syunyaev 2025, WP)
  • Observability: visible facts (e.g., economy, disaster management) constrain narrative power (Rosenfeld 2018, AJPS; Chang et al. 2022, PNAS)
  • Audience: controlled information environments create both forced and voluntary exposure; even educated/skeptical can be moved (Syunyaev 2025, WP)
  • Backfire: crude propaganda alienates moderates and opponents; sudden blocks spur circumvention (Huang 2018, JOP; Hobbs & Roberts 2018, APSR)
  • Access and state control: search ordering, removals, flooding (Roberts 2018, Princeton; King, Pan & Roberts 2017, APSR; Stukal et al. 2017, Big Data); People are interested in entertainment \(\Rightarrow\) incidental exposure to propaganda (Simonov & Rao 2022, JPE)

So what can we do?


  • Media literacy: reduces misperceptions in controlled settings (Badrinathan 2021, APSR) but could not be scalable and hard to assess

  • Critical thinking: corrections help; analytic thinking improves discernment (but could be short-lived) (Nyhan et al. 2021, PNAS; Erlich et al. 2023, PolPsych; Shirikov & Syunyaev 2025, WP)

  • Alternative information: independent TV/radio shifts attitudes and votes (Enikolopov et al. 2011, AER; Gagliarducci, Nannicini & Paserman 2020, AEJ:Applied)

  • Nature can help too? crises trigger circumvention gateways (Chang et al. 2022, PNAS)

Design matters

(Blair, Gottlieb, Nyhan et al. 2024, COP)

Implications for democracies



  • Patterns rhyme: slanted media persuades and polarizes.

    • Fox News entry increased GOP vote share (DellaVigna & Kaplan 2007, QJE)
    • Cable positions magnify persuasion/polarization (Martin & Yurukoglu 2017, AER)

  • But institutions and plural media markets blunt single narrative dominance (Guess 2020, BJPS; Guess 2021, AJPS) \(\Rightarrow\) social networks and self-selection are more important!

  • Summary: mechanisms overlap; constraints and intensity differ and complicate the study

Takeaways & open questions

  • Propaganda (and misinformation) can be effective, but the effects are not as clear-cut

    • Effects are likely conditional on content, audience, environment and population level effects can be exaggerated

  • We know little about what exactly works about propaganda \(\Rightarrow\) not clear how to counter it (both design and targeting)

  • Open questions:

    • What are the tools used by propaganda and why governments (political actors) choose them?
    • What are the implications of expansion of internet and digital tools, especially AI? (can be used for both propaganda/misinformation but also for countering it)
    • Are propaganda effects durable or they also diminish if not reinforced?
    • What leads to long-run habit change? Could be education and income, partisanship, social networks, critical thinking